Effect of interhospital transfer patterns on hospitals' quality performance measurement results

Abstract

Listen

Translate article

Study objectives: National measures for quality of care (QOC) for acute myocardial infarction (AMI) are emerging, and results of these measures are used to communicate and compare hospitals' QOC. Although the AMI QOC measures use the ideal candidate methodology (theoretically immune to differences in patient cohort characteristics), underuse among ideal candidates is not independent of patient characteristics. Interhospital transfer patterns are likely to accentuate differences in cohort characteristics at different hospitals, and these cohort differences may lead to loss of comparability of QOC measures across hospitals. We study the difference in cohort characteristics and QOC results between transfer-out (T) hospitals and receiving (R) hospitals in Connecticut (CT). Methods: Our analysis cohort (n=1,356) was assembled from AMI QOC assessments by the CT Quality Improvement Organization in 1998 and 2001. We excluded patients younger than 65 years. Hospitals were divided into those that transferred out less than 10% (R) and those that transferred out 10% or more of their AMI patients. We compared AMI cohort characteristics and AMI QOC for administration to ideal candidates of aspirin and β-blocker within the first day after hospitaccording to transfer pattern of the hospitals and transfer status of the cases. Results: Twenty-two T hospitals cared for 683 AMI cases, whereas 7 R hospitals cared for 673 AMI cases. T hospitals transferred out 221 (32.4%) of their AMI cases, and 271 (40.3%) were received in transfer to R hospitals. Cases transferred out of T hospitals or into R hospitals were younger, less often female patients, with fewer comorbidities than cases not transferred (Table). QOC measures were higher for cases transferred (Table), particularly for T hospitals, which outperformed R hospitals on early QOC ( P P Conclusion: Current QOC measures disadvantage hospitals that transfer out a substantial portion of AMI cases to tertiary care centers.Table, abstract 374.Transfer Hospitals (683 Cases)Receiving Hospitals (673 Cases)Not TransferredTransferred OutTransferred InNot TransferredMean age, y (SD)82 (8)76 (6)75 (7)78 (8)Female sex, %60424554Frail, %336719Previous stroke, %22121115Diabetes, %36252633Creatinine >2.5 mg/dL, %2581118Heart failure on admission, %44322144QOCAspirin on admission, %8894NA87β-blocker on admission, %7082NA66Aspirin at discharge, %85NA87 ∗ 88 ∗ β-blocker at discharge, %73NA92 ∗ 80 ∗